粒子群优化小波神经网络在惯导系统传递对准中的应用

为克服系统阶次较高时卡尔曼滤波实时性较差的特点,利用小波神经网络具有收敛速度快,结构简单,计算量小等优点,将其应用于惯导系统传递对准。并采用粒子群算法优化小波神经网络参数,取代传统的梯度下降法,克服了网络参数选取的盲目性,避免陷入局部极小值。仿真结果表明采用粒子群优化的小波神经网络进行传递对准,不仅获得了与Kalman滤波相同的精度,而且收敛速度快,满足了对准的实时性要求。

Application of Wavelet Neural Network Based on PSO Algorithm to Transfer Alignment of INS

To overcome the shortage of Kalman filter's bad real time quality when system step was high, wavelet neural network (WNN) was applied to transfer alignment of INS because of its easy structure and less calculation. The particle swarm optimization (PSO) algorithm was used to train the parameters of the WNN instead of the stochastic gradient algorithms, overcoming the blindness of parameter selection, avoiding trapping in the local minimal. Simulation results show that using WNN based on PSO to transfer alignment, not only get the precision similar to Kaiman filter, but also accelerate the rapidity of convergence and satisfy the real-time quality of transfer alignment.